Quantum Data for Quantum Machine Learning

This project explores how quantum machine learning can be used to learn from quantum data: data generated by quantum states, quantum circuits, quantum processors, simulators, sensors, and future quantum networks. Unlike standard machine learning, which usually works with classical datasets such as images, text, or numerical records, this project focuses on data whose structure is governed by quantum mechanics. Such data may appear as measurement outcomes, correlations, quantum state samples, circuit-generated distributions, or information produced by noisy quantum devices.

The aim is to develop practical and theoretically grounded methods for extracting useful information from quantum systems. This includes designing and benchmarking quantum machine learning models that can classify quantum states, recognize quantum phases, predict physical properties, detect anomalies in quantum measurement streams, and learn compact representations of many-body or device-generated quantum data. The project will study approaches such as quantum feature maps, variational quantum circuits, quantum kernel methods, and hybrid quantum-classical learning models.

A key challenge is that quantum data is expensive and difficult to access. Measurements are limited, quantum states may be costly to prepare, and present-day devices are affected by noise, finite sampling, and hardware constraints. Therefore, the project will place strong emphasis on data-efficient learning, using tools such as classical shadows, adaptive measurements, compressed representations, and physically motivated circuit ansätze to reduce measurement overhead and improve reliability.

The expected outcomes include new algorithms, benchmark datasets, theoretical analyses, and software tools for quantum-data-aware machine learning. The project will also produce demonstrator applications across quantum computing, quantum communication, and quantum sensing. By connecting quantum information science with modern machine learning, this project will support QC2’s mission to develop computational tools for next-generation quantum technologies and train students in quantum algorithms, quantum data analysis, and hybrid quantum-classical methods.